CN108319926A - A kind of the safety cap wearing detecting system and detection method of building-site - Google Patents
A kind of the safety cap wearing detecting system and detection method of building-site Download PDFInfo
- Publication number
- CN108319926A CN108319926A CN201810143710.8A CN201810143710A CN108319926A CN 108319926 A CN108319926 A CN 108319926A CN 201810143710 A CN201810143710 A CN 201810143710A CN 108319926 A CN108319926 A CN 108319926A
- Authority
- CN
- China
- Prior art keywords
- image
- safety cap
- detection
- building
- site
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/40—Scenes; Scene-specific elements in video content
- G06V20/46—Extracting features or characteristics from the video content, e.g. video fingerprints, representative shots or key frames
- G06V20/47—Detecting features for summarising video content
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/52—Surveillance or monitoring of activities, e.g. for recognising suspicious objects
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N5/00—Details of television systems
- H04N5/14—Picture signal circuitry for video frequency region
- H04N5/144—Movement detection
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N7/00—Television systems
- H04N7/18—Closed-circuit television [CCTV] systems, i.e. systems in which the video signal is not broadcast
- H04N7/181—Closed-circuit television [CCTV] systems, i.e. systems in which the video signal is not broadcast for receiving images from a plurality of remote sources
Landscapes
- Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Signal Processing (AREA)
- Data Mining & Analysis (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Cheminformatics (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Image Analysis (AREA)
- Alarm Systems (AREA)
Abstract
The invention discloses a kind of safety caps of building-site to wear detecting system and detection method, belongs to video identification detection field.For easy misrecognition existing in the prior art, the problem that safety cap data demand is more, recognition effect is bad, the present invention provides a kind of safety caps of building-site to wear detecting system and detection method, video-unit and detecting system, by A, acquisition video monitoring image, detection identification model is established；B, the image of acquisition building site real time monitoring carries out moving object segmentation, and identifying system is uploaded to after being tested with the image of moving object；C, the image that identifying system receiving step B is sent is detected identification, judges safety cap state；D, result is transmitted in message queue；E, detection recognition result is shown.It may be implemented accurate judgement safety cap and whether there is, it is desirable that initial data is few, and identification is accurate.
Description
Technical field
The present invention relates to video identification detection fields, and inspection is worn more specifically to a kind of safety cap of building-site
Examining system and detection method.
Background technology
Safety cap be it is a kind of prevent alluvium injury head protective articles need to pass through pendant into the personnel of construction site
It wears a safety helmet, to protect head, prevents the injury of accidental falling object.But it often has into construction site personnel and does not wear correctly
The case where wearing a safety helmet monitors the construction site in real time preferably to avoid such case, and detection is
The no detection according to regulation safe wearing cap is most important.By installing camera in inlet, regarded to what video camera took
Frequency is handled in real time, to judge worker's safe wearing cap situation and timely warning reminding, can effectively reduce artificial monitoring
Labour, and the omission factor manually monitored can be reduced.With the continuous development of computer vision technique, more and more fields
Conjunction begins to use the safety detection technology based on video analysis.Safety cap detection based on video analysis also possesses some special knowledge.
By retrieval, Chinese patent application, application number 201610641223.5, discloses publication date on January 4th, 2017
A kind of personnel safety cap wear condition real-time detection method based on video analysis, it is first, real from the camera in monitoring region
When read collected video frame, then classify the imagery exploitation detector of present frame to identify upper half of human body, will
The upper half of human body peak position that the target detected is positioned and confirmly detected, the point extraction phase on the basis of peak
The region answered, the color of the image-region is finally judged according to color characteristic to judge to be detected personnel whether safe wearing
Cap, if finding, detection zone has the non-safe wearing cap of personnel, carries out alarm to ensure the peace into working region personnel
Entirely.The invention can reduce the influence that background environment detects safety cap, and high degree improves whether identification region personnel wear
The accuracy to wear a safety helmet, while having that detection speed is fast, the good feature of real-time.But this scheme is based on human figure and peace
The detection identification of full cap color, lacks robustness, does not adapt to complex environment and the changeable form and color of cap.Example
Such as, common hat has the color of safety cap, and such case cap can be misidentified.
In addition, Chinese patent application, application number 201410311121.8, publication date September in 2014 10 days discloses one kind
Construction site safety cap wear condition monitoring method, hardware used in the monitoring method include at least：High-speed ball camera, prison
Control device, display device；The monitoring method includes at least：The system of the acquisition of monitoring image, the processing of monitoring image and data
Meter and display；The acquisition of monitoring image includes that control video camera is continuously shot；The processing of monitoring image includes：The mark of target
The identification of note and safety cap；The label of target includes：Obtain the current foreground of binaryzation, label and the commissarial movement mesh of tracking
Mark；The label of detection and identification based on target of safety cap, and the statistics of data and display are then according to the detection of safety cap and knowledge
Is not counted and calculated；Wearing of the monitoring method of the present invention by vision-based inspection method to construction site inlet
The case where safety cap, is monitored, which can realize automatic detection and monitoring.But the rule-based monitoring of this scheme
System not over machine learning or deep learning identification human body and object, therefore lacks necessary robustness, cannot be good
The safety cap for adapting to complicated building site environment monitors demand.
In addition, patent application 201610778655.0, publication date on 2 1st, 2017, provide what a kind of safety cap was worn
Detection method and device, camera, server, including：Live video image is obtained, the position of human body established by training study
Detection model is detected to obtain position of human body to the live video image, is judged whether in safety according to the position of human body
Cap wears region；If it is, the number of people safety cap joint-detection model established by training study is to the live video figure
As being detected to obtain number of people safety cap united state, safe wearing cap is judged whether according to number of people safety cap united state, such as
Fruit does not have safe wearing cap, then safety cap is worn not verified；If safe wearing cap, pass through the peace of training study foundation
Full cap type detection model is detected the live video image to obtain safety cap type, judges whether safety cap type accords with
It closes and provides, if meeting regulation, safety cap is worn through verification, and otherwise, safety cap wears not verified, raising safety cap
The accuracy of the testing result of wearing.Being detected based on the safety cap of machine learning and depth learning technology for this scheme is identified, is lacked
Weary matched cloud platform and highly effective algorithm framework, because the complicated building site that system is not suitable for multi-cam monitors demand.It should
System does not support the front end of persisted moniker safety cap sample, can not be by the sample and deep learning algorithm continuous learning that newly mark
Identification model is detected, to influence actual detection recognition effect.
Invention content
1. technical problems to be solved
For easy misrecognition existing in the prior art, the problem that safety cap data demand is more, recognition effect is bad, originally
Invention provides a kind of the safety cap wearing detecting system and detection method of building-site, and accurate judgement safety cap may be implemented in it
It whether there is, it is desirable that initial data is few, and identification is accurate.
2. technical solution
The purpose of the present invention is achieved through the following technical solutions.
A kind of safety cap of building-site wears detecting system, including video-unit and detecting system, video-unit are
Camera, the detecting system include the video stream access module set gradually, image capture module, movement human detection mould
Block, image uploading module, key images extraction module, safety cap detection module；Testing result is sent by safety cap detection module
Message subscribing/release module and memory module.
Further, memory module preserves testing result and original image library.
Further, image is uploaded to original image library and key images extraction module by image uploading module.
A kind of safety cap wearing detection method of building-site, steps are as follows：
A, video monitoring image is acquired, detection identification model is established；
B, the image of acquisition building site real time monitoring carries out moving object segmentation, is uploaded after being tested with the image of moving object
To identifying system；
C, the image that identifying system receiving step B is sent is detected identification, judges safety cap state；
D, result is transmitted in message queue；
E, detection recognition result is shown.
Further, step A is as follows：
A1, acquisition building-site video monitoring image, manually mark out detection and identify required picture material and format；
A2, identification model is detected using the picture training deep learning SSD manually marked；
A3, the data set test detection identification model different from training data, detection recognition result are used.
Further, in step A2, the Detection tasks of SSD detection identification models includes proposal layers, recurrence task,
Classification task, a proposal is corresponded to there are four output in recurrence task, and classification task needs to carry out each proposal
Classification, wherein proposal layers includes N number of proposal, and N is natural number, and the output of the recurrence branch of corresponding recurrence task is
4N, for branch of classifying, training M classes, M is natural number, and the number of output for branch of classifying is 4 (M+1), and M+1 indicates concrete class
In addition a kind of background.
Further, in step A2, it includes focused lost function FL and cross entropy loss function that SSD, which detects identification model,
CE, wherein
FL(pt)=- (1-pt)γlog(pt)
CE(pt)=- log (pt)
Pt is the probability of safety cap classification, and pt is variable, and ranging from 0-1, γ are focused lost parameters, and γ is parameter, γ
=2.
Further, further include following steps in step A3, recycled and carried out using the data set different from training data
Step A2, optimizing detection recognition result obtain more accurate identification model.
Further, in step B, Image Acquisition judgement is as follows：.
B1, pass through camera, real-time monitoring building-site video flow monitoring；
B2, Image Acquisition is carried out to video flow monitoring, judges whether abnormal conditions occur according to connection return value, if occurring
It is abnormal, realize that Image Acquisition service is restarted automatically；
B3, setting image file block queue, and three image scenes are often acquired in queue and carry out moving object segmentation, are judged
Whether there is moving object in picture；
If B4, without moving object, return continues to acquire image；
If B5, having detected that picture is uploaded to identifying system by moving object, request.
Further, in step C, safety cap detection identification is as follows：.
C1, the moving object image that step B is uploaded is converted to critical data picture as the input of identification service；
C2, with the deep learning SSD model inspections based on focused lost function identify in the picture whether someone；
If C3, nobody, unmanned state's code is returned, coordinate is not returned；
If C4, someone, with the deep learning SSD model inspections based on focused lost function identify the people whether safe wearing
Cap,；
If C5, there is no safe wearing cap, no safe wearing cap conditional code is returned, the coordinate of people is returned；
If C6, safe wearing cap, safe wearing cap conditional code is returned, the coordinate of people is returned, makes corresponding output.
3. advantageous effect
Compared with the prior art, the advantage of the invention is that：
(1) present invention can be adapted to the video flowing of access by VAM Video Access Module, have better Function Extension
Property and more extensive video equipment adaptability；
(2) the moving object segmentation model that the present invention is established by machine learning, in front end, front end processor is to moving object wheel
Wide and position is detected, and the video frame information that need to be detected is positioned and marked；
(3) video frame extraction and the compression of the invention to detecting moving object, reduces the data volume for uploading high in the clouds, improves
Communication efficiency reduces network cost and requirement；
(4) single more frame detection framework (ssd) depth of the invention by being based on focused lost function (focalloss)
It practises human testing and safety cap detects recognizer, beyond the clouds video frame key images information of the detection identification server to upload
It is detected identification, critical body points and safety cap can be accurately positioned from complicated video；
(5) present invention can be met using the detection identification module of full deep learning under gpu supports with large-scale parallel
The multi-cam demand of actual scene；
(6) invention supports the front-end interface of persisted moniker safety cap sample, shell to pass through the sample and deep learning that newly mark
Algorithm continuous learning detects identification model, to be disposed in more scenes；
(7) present invention can widely utilize the result of detection identification, example by news release/subscribing module
As and front-end interface communication, with better autgmentability.
Description of the drawings
Fig. 1 is the system structure diagram of the present invention；
Fig. 2 is the overhaul flow chart of the present invention；
Fig. 3 is the Image Acquisition service flow diagram of the present invention；
Fig. 4 is safety cap detection identification service flow diagram.
Specific implementation mode
With reference to the accompanying drawings of the specification and specific embodiment, the present invention is described in detail.
Embodiment 1
This programme detected automatically by computer system site staff whether safe wearing cap, for do not wear a safety helmet disobey
Rule behavior, triggering alarm and recording-related information.
As shown in Figure 1, the safety cap that this programme is provided with building-site wears detecting system, including video-unit and inspection
Examining system, video-unit are camera, and camera can be arranged as required to several, and setting is being entered and building-site work
At artificial work, detecting system is background system, and the detecting system includes the video stream access module set gradually, can be docked
The video flowing entered is adapted to, and has better function expansibility and more extensive video equipment adaptability；Image capture module,
Front end front end processor is detected moving object contours and position, and the video frame information that need to be detected is positioned and marked, this
The label at place can be by being manually marked；Movement human detection module, image uploading module, key images extract mould
Block, the video frame extraction to moving object and compression, reduce upload high in the clouds data volume, improve communication efficiency, reduce network at
Originally it and requires；Safety cap detection module；Testing result is sent into message subscribing/release module and storage mould by safety cap detection module
Block, news release/subscribing module can widely utilize the result of detection identification, such as the communication with front-end interface,
With better autgmentability, memory module preserves testing result and original image library.Image is uploaded to original by image uploading module
Beginning image library and key images extraction module.
Embodiment 2
As shown in Figure 2,3, 4, the system based on above-described embodiment 1, safety cap wear detection method, and steps are as follows：
A, video monitoring image is acquired, detection is manually marked out and identifies required picture material and format, it is artificial to mark
There are head zones of safe wearing cap etc. in the dated video pictures of palpus, are labeled according to demand to color, safety cap type,
Establish detection identification model；Video flowing can be obtained by accessing the video file of video camera or storage herein, be used herein as
The picture training deep learning SSD detection identification models manually marked, the Detection task of SSD detection identification models include
Proposal layers, return task, classification task, a proposal is corresponded to there are four output in recurrence task, and classification task needs
To classify to each proposal, wherein proposal layers includes N number of proposal, and N is natural number, and corresponding return is appointed
The output of the recurrence branch of business is 4N, and for branch of classifying, training M classes, M is natural number, and the number of output for branch of classifying is 4 (M
+ 1), M+1 indicates concrete class plus a kind of background.Therein 1 indicates background class branch.According to this thinking, with mark
The sample of note, so that it may to design the net web frame of safety cap needs, carry out the training of model.
In the present solution, having used focused lost function in the identification of SSD model safety caps, the intersection entropy loss of standard is improved
Function ensure that our model can detect identification not well for solving the problems, such as that safety cap sample is less in practice
It is below the concrete form of focused lost function FL and cross entropy loss function CE, wherein p with the safety cap of form and colort
For the probability of safety cap classification, γ is focused lost parameter.The exponential factor of wherein FL first halfs is focused lost and intersection
Where the difference of entropy loss；ptFor variable, ranging from 0-1, γ is preset parameter, γ=2.
FL(pt)=- (1-pt)γlog(pt)
CE(pt)=- log (pt)
Using the data set test detection identification model different from training data, recognition result is detected, and by above-mentioned
Optimization order detects recognition result, and model training is detached with test data, ensures that effect can be extensive reproducible, model itself passes through
The data newly marked can be also iterated.
B, the image of acquisition building site real time monitoring carries out moving object segmentation, is uploaded after being tested with the image of moving object
To identifying system；
It is as follows：B1, pass through camera, real-time monitoring building-site video flow monitoring；Monitoring stream herein, this
Scheme uses RTSP video flow monitorings；A cycle can be set, such as obtains a frame video image every a cycle (1s), is used
It is identified in detection；
B2, Image Acquisition is carried out to RTSP video flow monitorings based on Ffmpeg, abnormal conditions is judged according to connection return value
Whether occur, if there is exception, realizes that Image Acquisition service is restarted automatically；
B3, setting image file block queue, and three image scenes are often acquired in queue and carry out moving object segmentation, are judged
Whether there is moving object in picture；It often acquires three image scenes and does the once moving object segmentation based on machine learning algorithm,
Machine learning algorithm herein can use judgment method in the prior art, such as vector machine, convolutional neural networks, herein not
It summarizes, by the way that video background content is established the probabilistic model based on gauss hybrid models, utilizes the information dynamic of three figures
The probabilistic model for updating background, and judges whether there is moving object in picture, in practice the selection of three figures well adapt to whole
The rate that body calculates, it is of course also possible to be judged according to demand using more than three pictures.
If B4, without moving object, return continues to acquire image；
If B5, having detected moving object, detected video image is cached, generates corresponding information collection label, and
Original image library is uploaded after human body contour outline is marked, and picture is uploaded to by identifying system by HTTP Post requests.
C, the image that identifying system receiving step B is sent is detected identification, judges safety cap state；Step C is specifically walked
It is rapid as follows：
C1, the moving object image that step B is uploaded is converted to critical data picture as the input of identification service；Herein
Base64 pictures can be converted into as critical data picture, the pixel of human body contour outline and surrounding in buffered video image is extracted and believe
Breath generates image to be detected file；
C2, with the deep learning SSD model inspections based on focused lost function identify in the picture whether someone；
If C3, nobody, unmanned state's code is returned to, it is unmanned state's code that conditional code 2 can such as be arranged herein, does not return to seat
Mark；
If C4, someone, with the deep learning SSD model inspections based on focused lost function identify the people whether safe wearing
Cap can set a p hereintThreshold value, threshold range 0.6-0.8, the present embodiment select 0.7 be used as threshold value；
If C5, not having safe wearing cap, no safe wearing cap conditional code is returned to, conditional code 1, which such as can be arranged, herein is
Unmanned state's code, returns to the coordinate of people；
If C6, safe wearing cap, safe wearing cap conditional code is returned to, it is unmanned state that conditional code 0 can such as be arranged herein
Code, returns to the coordinate of people, makes corresponding output.
D, result is transmitted in message queue；
E, detection recognition result is shown.The region set by terminal, background system etc. can be shown to.
Schematically the invention and embodiments thereof are described above, description is not limiting, not
In the case of the spirit or essential characteristics of the present invention, the present invention can be realized in other specific forms.Institute in attached drawing
What is shown is also one of the embodiment of the invention, and actual structure is not limited to this, any attached in claim
Icon note should not limit the claims involved.So if those skilled in the art are enlightened by it, do not departing from
In the case of this creation objective, frame mode similar with the technical solution and embodiment are not inventively designed, it should all
Belong to the protection domain of this patent.In addition, one word of " comprising " is not excluded for other elements or step, "one" word before the component
It is not excluded for including " multiple " element.The multiple element stated in claim to a product can also by an element by software or
Person hardware is realized.The first, the second equal words are used to indicate names, and are not represented any particular order.
Claims (10)
1. a kind of safety cap of building-site wears detecting system, including video-unit and detecting system, video-unit is to take the photograph
As head, it is characterised in that：The detecting system includes the video stream access module set gradually, image capture module, movement
Human detection module, image uploading module, key images extraction module, safety cap detection module；Safety cap detection module will be examined
It surveys result and is sent into message subscribing/release module and memory module.
2. a kind of safety cap of building-site according to claim 1 wears detecting system, it is characterised in that：Memory module
Preserve testing result and original image library.
3. a kind of safety cap of building-site according to claim 1 wears detecting system, it is characterised in that：Image uploads
Image is uploaded to original image library and key images extraction module by module.
4. a kind of safety cap of building-site wears detection method, steps are as follows：
A, video monitoring image is acquired, detection identification model is established；
B, the image of acquisition building site real time monitoring carries out moving object segmentation, and knowledge is uploaded to after being tested with the image of moving object
Other system；
C, the image that identifying system receiving step B is sent is detected identification, judges safety cap state；
D, result is transmitted in message queue；
E, detection recognition result is shown.
5. a kind of safety cap of building-site according to claim 4 wears detection method, it is characterised in that：Step A tools
Steps are as follows for body：
A1, acquisition building-site video monitoring image, manually mark out detection and identify required picture material and format；
A2, identification model is detected using the picture training deep learning SSD manually marked；
A3, the data set test detection identification model different from training data, detection recognition result are used.
6. a kind of safety cap of building-site according to claim 5 wears detection method, it is characterised in that：Step A2
In, the Detection tasks of SSD detection identification models includes proposal layers, recurrence task, classification task, one in recurrence task
Proposal is corresponded to there are four output, and classification task needs are classified to each proposal, wrapped in wherein proposal layers
N number of proposal is included, N is natural number, and the output of the recurrence branch of corresponding recurrence task is 4N, for branch of classifying, training M
Class, M are natural number, and the number of output for branch of classifying is 4 (M+1), and M+1 indicates concrete class plus a kind of background.
7. a kind of safety cap of building-site according to claim 5 or 6 wears detection method, it is characterised in that：Step
In A2, SSD detection identification models include focused lost function FL and cross entropy loss function CE, wherein
FL(pt)=- (1-pt)γlog(pt)
CE(pt)=- log (pt)
Pt is the probability of safety cap classification, and pt is variable, and ranging from 0-1, γ are focused lost parameters, and γ is parameter, γ=2.
8. a kind of safety cap of building-site according to claim 5 wears detection method, it is characterised in that：In step A3
Further include following steps, step A2, optimizing detection recognition result are carried out using the data set cycle different from training data.
9. a kind of safety cap of building-site according to claim 4 or 5 wears detection method, it is characterised in that：
In step B, Image Acquisition judgement is as follows：
B1, pass through camera, real-time monitoring building-site video flow monitoring；
B2, Image Acquisition is carried out to video flow monitoring, judges whether abnormal conditions occur according to connection return value, if occurring different
Often, realize that Image Acquisition service is restarted automatically；
B3, setting image file block queue, and three image scenes are often acquired in queue and carry out moving object segmentation, judge picture
In whether have moving object；
If B4, without moving object, return continues to acquire image；
If B5, having detected that picture is uploaded to identifying system by moving object, request.
10. a kind of safety cap of building-site according to claim 9 wears detection method, it is characterised in that：Step C
In, safety cap detection identification is as follows：.
C1, the moving object image that step B is uploaded is converted to critical data picture as the input of identification service；
C2, with the deep learning SSD model inspections based on focused lost function identify in the picture whether someone；
If C3, nobody, unmanned state's code is returned, coordinate is not returned；
If C4, someone, with the deep learning SSD model inspections based on focused lost function identify the people whether safe wearing cap,；
If C5, there is no safe wearing cap, no safe wearing cap conditional code is returned, the coordinate of people is returned；
If C6, safe wearing cap, safe wearing cap conditional code is returned, the coordinate of people is returned, makes corresponding output.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810143710.8A CN108319926A (en) | 2018-02-12 | 2018-02-12 | A kind of the safety cap wearing detecting system and detection method of building-site |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810143710.8A CN108319926A (en) | 2018-02-12 | 2018-02-12 | A kind of the safety cap wearing detecting system and detection method of building-site |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108319926A true CN108319926A (en) | 2018-07-24 |
Family
ID=62904019
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810143710.8A Pending CN108319926A (en) | 2018-02-12 | 2018-02-12 | A kind of the safety cap wearing detecting system and detection method of building-site |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108319926A (en) |
Cited By (32)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108985276A (en) * | 2018-08-21 | 2018-12-11 | 盯盯拍（深圳）技术股份有限公司 | Vision AI algorithmic system design method, vision AI algorithmic system design device and vision AI algorithmic system |
| CN109255298A (en) * | 2018-08-07 | 2019-01-22 | 南京工业大学 | Safety cap detection method and system in a kind of dynamic background |
| CN109447479A (en) * | 2018-10-31 | 2019-03-08 | 广西路桥工程集团有限公司 | One kind being based on three-dimensional visualization project management system |
| CN109558783A (en) * | 2018-09-20 | 2019-04-02 | 中建科技有限公司深圳分公司 | A kind of violation detection method, system and equipment for construction site |
| CN110070033A (en) * | 2019-04-19 | 2019-07-30 | 山东大学 | Safety cap wearing state detection method in a kind of power domain dangerous work region |
| CN110084253A (en) * | 2019-05-05 | 2019-08-02 | 厦门美图之家科技有限公司 | A method of generating object detection model |
| CN110188724A (en) * | 2019-06-05 | 2019-08-30 | 中冶赛迪重庆信息技术有限公司 | The method and system of safety cap positioning and color identification based on deep learning |
| CN110222672A (en) * | 2019-06-19 | 2019-09-10 | 广东工业大学 | The safety cap of construction site wears detection method, device, equipment and storage medium |
| CN110263686A (en) * | 2019-06-06 | 2019-09-20 | 温州大学 | A kind of construction site safety of image cap detection method based on deep learning |
| CN110618436A (en) * | 2019-04-04 | 2019-12-27 | 中国石油大学(北京) | Inspection method, device and equipment based on instant positioning and map construction |
| CN110728264A (en) * | 2019-10-24 | 2020-01-24 | 名创优品（横琴）企业管理有限公司 | Wearing condition detection method and system for accessory wearable product |
| CN110929608A (en) * | 2019-11-12 | 2020-03-27 | 国网天津市电力公司 | Safety helmet wearing detection method and system |
| CN110956823A (en) * | 2020-02-21 | 2020-04-03 | 浩鲸云计算科技股份有限公司 | Traffic jam detection method based on video analysis |
| CN111062303A (en) * | 2019-12-12 | 2020-04-24 | 深圳市捷顺科技实业股份有限公司 | Image processing method, system and computer storage medium |
| CN111083441A (en) * | 2019-12-18 | 2020-04-28 | 广州穗能通能源科技有限责任公司 | Construction site monitoring method and device, computer equipment and storage medium |
| CN111191586A (en) * | 2019-12-30 | 2020-05-22 | 安徽小眯当家信息技术有限公司 | Method and system for inspecting wearing condition of safety helmet of personnel in construction site |
| CN111476083A (en) * | 2020-02-07 | 2020-07-31 | 山东理工大学 | Automatic identification method for wearing of safety helmet of electric power staff |
| CN111639546A (en) * | 2020-05-07 | 2020-09-08 | 金钱猫科技股份有限公司 | Small-scale target cloud computing identification method and device based on neural network |
| CN111814762A (en) * | 2020-08-24 | 2020-10-23 | 深延科技(北京)有限公司 | Helmet wearing detection method and device |
| CN111860181A (en) * | 2020-06-23 | 2020-10-30 | 国网山东省电力公司淄博供电公司 | Distributed safety helmet identification system |
| CN111914636A (en) * | 2019-11-25 | 2020-11-10 | 南京桂瑞得信息科技有限公司 | Method and device for detecting whether pedestrian wears safety helmet |
| CN112184773A (en) * | 2020-09-30 | 2021-01-05 | 华中科技大学 | Helmet wearing detection method and system based on deep learning |
| CN112188164A (en) * | 2020-09-29 | 2021-01-05 | 爱动超越人工智能科技（北京）有限责任公司 | AI vision-based violation real-time monitoring system and method |
| CN112257620A (en) * | 2020-10-27 | 2021-01-22 | 广州华微明天软件技术有限公司 | Safe wearing condition identification method |
| CN112347824A (en) * | 2019-08-09 | 2021-02-09 | 中移信息技术有限公司 | Wearing object identification method, device, equipment and storage medium |
| CN112466086A (en) * | 2020-10-26 | 2021-03-09 | 福州微猪信息科技有限公司 | Visual identification early warning device and method for farm work clothes |
| CN112949354A (en) * | 2019-12-10 | 2021-06-11 | 顺丰科技有限公司 | Method and device for detecting wearing of safety helmet, electronic equipment and computer-readable storage medium |
| CN113033492A (en) * | 2021-04-25 | 2021-06-25 | 内蒙古海洋工信科技有限责任公司 | Magnetic material preparation process identification and monitoring system based on artificial intelligence |
| CN114285874A (en) * | 2021-12-07 | 2022-04-05 | 重庆邮电大学 | Safety helmet timer scheduling method and system based on abnormal uncapping record |
| CN115086778A (en) * | 2021-03-15 | 2022-09-20 | 中移(上海)信息通信科技有限公司 | AI processing method and device for video stream |
| CN115731670A (en) * | 2022-11-15 | 2023-03-03 | 广州一洲信息技术有限公司 | Safety alarm method, device and monitoring system for construction site |
| CN115086778B (en) * | 2021-03-15 | 2024-05-24 | 中移(上海)信息通信科技有限公司 | AI processing method and device for video stream |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104021648A (en) * | 2014-06-22 | 2014-09-03 | 朱明增 | Intelligent monitoring and early warning system for safety helmet wearing on transformer substation site |
| CN104063722A (en) * | 2014-07-15 | 2014-09-24 | 国家电网公司 | Safety helmet identification method integrating HOG human body target detection and SVM classifier |
| CN104899559A (en) * | 2015-05-25 | 2015-09-09 | 江苏大学 | Rapid pedestrian detection method based on video monitoring |
| CN106372662A (en) * | 2016-08-30 | 2017-02-01 | 腾讯科技（深圳）有限公司 | Helmet wearing detection method and device, camera, and server |
| CN106408000A (en) * | 2016-08-25 | 2017-02-15 | 国网浙江省电力公司杭州供电公司 | Method and device for intelligent detection on personnel safety helmet |
-
2018
- 2018-02-12 CN CN201810143710.8A patent/CN108319926A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104021648A (en) * | 2014-06-22 | 2014-09-03 | 朱明增 | Intelligent monitoring and early warning system for safety helmet wearing on transformer substation site |
| CN104063722A (en) * | 2014-07-15 | 2014-09-24 | 国家电网公司 | Safety helmet identification method integrating HOG human body target detection and SVM classifier |
| CN104899559A (en) * | 2015-05-25 | 2015-09-09 | 江苏大学 | Rapid pedestrian detection method based on video monitoring |
| CN106408000A (en) * | 2016-08-25 | 2017-02-15 | 国网浙江省电力公司杭州供电公司 | Method and device for intelligent detection on personnel safety helmet |
| CN106372662A (en) * | 2016-08-30 | 2017-02-01 | 腾讯科技（深圳）有限公司 | Helmet wearing detection method and device, camera, and server |
Non-Patent Citations (2)
| Title |
|---|
| TSUNG-YI LIN: "Focal Loss for Dense Object Detection", 《ARXIV》 * |
| WEI,LIU: "SSD: Single Shot MultiBox Detector", 《ARXIV》 * |
Cited By (39)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109255298A (en) * | 2018-08-07 | 2019-01-22 | 南京工业大学 | Safety cap detection method and system in a kind of dynamic background |
| CN108985276A (en) * | 2018-08-21 | 2018-12-11 | 盯盯拍（深圳）技术股份有限公司 | Vision AI algorithmic system design method, vision AI algorithmic system design device and vision AI algorithmic system |
| CN109558783A (en) * | 2018-09-20 | 2019-04-02 | 中建科技有限公司深圳分公司 | A kind of violation detection method, system and equipment for construction site |
| CN109447479A (en) * | 2018-10-31 | 2019-03-08 | 广西路桥工程集团有限公司 | One kind being based on three-dimensional visualization project management system |
| CN109447479B (en) * | 2018-10-31 | 2021-12-21 | 广西路桥工程集团有限公司 | Project management system based on three-dimensional visualization |
| CN110618436A (en) * | 2019-04-04 | 2019-12-27 | 中国石油大学(北京) | Inspection method, device and equipment based on instant positioning and map construction |
| CN110070033A (en) * | 2019-04-19 | 2019-07-30 | 山东大学 | Safety cap wearing state detection method in a kind of power domain dangerous work region |
| CN110070033B (en) * | 2019-04-19 | 2020-04-24 | 山东大学 | Method for detecting wearing state of safety helmet in dangerous working area in power field |
| CN110084253A (en) * | 2019-05-05 | 2019-08-02 | 厦门美图之家科技有限公司 | A method of generating object detection model |
| CN110188724A (en) * | 2019-06-05 | 2019-08-30 | 中冶赛迪重庆信息技术有限公司 | The method and system of safety cap positioning and color identification based on deep learning |
| CN110188724B (en) * | 2019-06-05 | 2023-02-28 | 中冶赛迪信息技术(重庆)有限公司 | Method and system for helmet positioning and color recognition based on deep learning |
| CN110263686A (en) * | 2019-06-06 | 2019-09-20 | 温州大学 | A kind of construction site safety of image cap detection method based on deep learning |
| CN110222672A (en) * | 2019-06-19 | 2019-09-10 | 广东工业大学 | The safety cap of construction site wears detection method, device, equipment and storage medium |
| CN112347824A (en) * | 2019-08-09 | 2021-02-09 | 中移信息技术有限公司 | Wearing object identification method, device, equipment and storage medium |
| CN110728264A (en) * | 2019-10-24 | 2020-01-24 | 名创优品（横琴）企业管理有限公司 | Wearing condition detection method and system for accessory wearable product |
| CN110929608A (en) * | 2019-11-12 | 2020-03-27 | 国网天津市电力公司 | Safety helmet wearing detection method and system |
| CN111914636A (en) * | 2019-11-25 | 2020-11-10 | 南京桂瑞得信息科技有限公司 | Method and device for detecting whether pedestrian wears safety helmet |
| CN112949354A (en) * | 2019-12-10 | 2021-06-11 | 顺丰科技有限公司 | Method and device for detecting wearing of safety helmet, electronic equipment and computer-readable storage medium |
| CN111062303A (en) * | 2019-12-12 | 2020-04-24 | 深圳市捷顺科技实业股份有限公司 | Image processing method, system and computer storage medium |
| CN111083441A (en) * | 2019-12-18 | 2020-04-28 | 广州穗能通能源科技有限责任公司 | Construction site monitoring method and device, computer equipment and storage medium |
| CN111191586A (en) * | 2019-12-30 | 2020-05-22 | 安徽小眯当家信息技术有限公司 | Method and system for inspecting wearing condition of safety helmet of personnel in construction site |
| CN111191586B (en) * | 2019-12-30 | 2020-11-27 | 安徽小眯当家信息技术有限公司 | Method and system for inspecting wearing condition of safety helmet of personnel in construction site |
| CN111476083A (en) * | 2020-02-07 | 2020-07-31 | 山东理工大学 | Automatic identification method for wearing of safety helmet of electric power staff |
| CN110956823A (en) * | 2020-02-21 | 2020-04-03 | 浩鲸云计算科技股份有限公司 | Traffic jam detection method based on video analysis |
| CN111639546A (en) * | 2020-05-07 | 2020-09-08 | 金钱猫科技股份有限公司 | Small-scale target cloud computing identification method and device based on neural network |
| CN111860181A (en) * | 2020-06-23 | 2020-10-30 | 国网山东省电力公司淄博供电公司 | Distributed safety helmet identification system |
| CN111814762A (en) * | 2020-08-24 | 2020-10-23 | 深延科技(北京)有限公司 | Helmet wearing detection method and device |
| CN112188164A (en) * | 2020-09-29 | 2021-01-05 | 爱动超越人工智能科技（北京）有限责任公司 | AI vision-based violation real-time monitoring system and method |
| CN112184773A (en) * | 2020-09-30 | 2021-01-05 | 华中科技大学 | Helmet wearing detection method and system based on deep learning |
| CN112466086A (en) * | 2020-10-26 | 2021-03-09 | 福州微猪信息科技有限公司 | Visual identification early warning device and method for farm work clothes |
| CN112257620B (en) * | 2020-10-27 | 2021-10-26 | 广州华微明天软件技术有限公司 | Safe wearing condition identification method |
| CN112257620A (en) * | 2020-10-27 | 2021-01-22 | 广州华微明天软件技术有限公司 | Safe wearing condition identification method |
| CN115086778A (en) * | 2021-03-15 | 2022-09-20 | 中移(上海)信息通信科技有限公司 | AI processing method and device for video stream |
| CN115086778B (en) * | 2021-03-15 | 2024-05-24 | 中移(上海)信息通信科技有限公司 | AI processing method and device for video stream |
| CN113033492A (en) * | 2021-04-25 | 2021-06-25 | 内蒙古海洋工信科技有限责任公司 | Magnetic material preparation process identification and monitoring system based on artificial intelligence |
| CN113033492B (en) * | 2021-04-25 | 2022-08-02 | 内蒙古海洋工信科技有限责任公司 | Magnetic material preparation process identification and monitoring system based on artificial intelligence |
| CN114285874A (en) * | 2021-12-07 | 2022-04-05 | 重庆邮电大学 | Safety helmet timer scheduling method and system based on abnormal uncapping record |
| CN114285874B (en) * | 2021-12-07 | 2023-11-03 | 中化学交通建设集团有限公司 | Safety helmet timer scheduling method and system based on abnormal uncapping record |
| CN115731670A (en) * | 2022-11-15 | 2023-03-03 | 广州一洲信息技术有限公司 | Safety alarm method, device and monitoring system for construction site |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108319926A (en) | A kind of the safety cap wearing detecting system and detection method of building-site | |
| CN110738127B (en) | Helmet identification method based on unsupervised deep learning neural network algorithm | |
| CN108062349B (en) | Video monitoring method and system based on video structured data and deep learning | |
| CN111598040B (en) | Construction worker identity recognition and safety helmet wearing detection method and system | |
| CN110443109A (en) | Abnormal behaviour monitor processing method, device, computer equipment and storage medium | |
| CN108090458A (en) | Tumble detection method for human body and device | |
| CN110516522B (en) | Inspection method and system | |
| KR102149832B1 (en) | Automated Violence Detecting System based on Deep Learning | |
| CN111881730A (en) | Wearing detection method for on-site safety helmet of thermal power plant | |
| KR102195706B1 (en) | Method and Apparatus for Detecting Intruder | |
| CN102811343A (en) | Intelligent video monitoring system based on behavior recognition | |
| CN112235537B (en) | Transformer substation field operation safety early warning method | |
| CN110728252B (en) | Face detection method applied to regional personnel motion trail monitoring | |
| CN111091098A (en) | Training method and detection method of detection model and related device | |
| CN112329691A (en) | Monitoring video analysis method and device, electronic equipment and storage medium | |
| CN107330414A (en) | Act of violence monitoring method | |
| CN111325133A (en) | Image processing system based on artificial intelligence recognition | |
| CN109635717A (en) | A kind of mining pedestrian detection method based on deep learning | |
| CN111860187A (en) | High-precision worn mask identification method and system | |
| CN116264004A (en) | Vision-based monitoring of field safety compliance based on worker re-identification and personal protective equipment classification | |
| CN106372566A (en) | Digital signage-based emergency evacuation system and method | |
| CN114067396A (en) | Vision learning-based digital management system and method for live-in project field test | |
| CN113052125B (en) | Construction site violation image recognition and alarm method | |
| CN105095891A (en) | Human face capturing method, device and system | |
| CN111723725A (en) | Multi-dimensional analysis system based on video AI |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| CB03 | Change of inventor or designer information |
Inventor after: Wang Shandong Inventor after: Yin Zhiwei Inventor after: Hua Zhikui Inventor before: Wang Shandong Inventor before: Yin Zhiwei Inventor before: Hua Zhikui Inventor before: Ding Lei Inventor before: Zhan Boqi Inventor before: Fan Zhiyuan |
|
| CB03 | Change of inventor or designer information | ||
| PB01 | Publication | ||
| PB01 | Publication | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20180705 Address after: 243000 the 6 floor of 2 Ma'anshan software garden, 698 Huo Li Shan Road, Hua Shan, Ma'anshan, Anhui. Applicant after: Anhui Jinhe software Limited by Share Ltd Address before: 243000 the 6 floor of 2 Ma'anshan software garden, 698 Huo Li Shan Road, Hua Shan, Ma'anshan, Anhui. Applicant before: Anhui Jinhe software Limited by Share Ltd Applicant before: Beijing Xiang Xiang Intelligent Technology Co., Ltd. |
|
| TA01 | Transfer of patent application right | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20180724 |
|
| RJ01 | Rejection of invention patent application after publication |